Imagine trying to teach a robot to manage a building, but you don't have enough real-world examples to train it. This project creates high-quality 'fake' data that looks and acts like the real thing to train AI safely. It also ensures the AI stays energy-efficient and easy to understand as it learns on the job.
AI for smart spaces often fails because real-world data is too expensive to collect, too sensitive to share, or too incomplete to train reliable models.
A system for generating synthetic training data and a set of energy-efficient AI services (AIaaS/CaaS) for the IoT-Edge-Cloud continuum.
If you are a factory operator dealing with gaps in sensor data for predictive maintenance — this project developed synthetic data generation and uncertainty quantification that allows AI models to be trained and tested without needing years of expensive historical data.
If you are a building manager dealing with high energy costs for AI systems — this project developed green operation techniques and energy-efficient continual learning that reduces the carbon footprint of your smart space AI.
If you are a grid provider dealing with sensitive data sharing restrictions — this project developed federated representation learning and trustworthy datasets that allow AI to improve without exposing private infrastructure data.
Based on available project data, specific pricing or implementation costs are not provided.
Yes, the project is validating its mechanisms through five pilot cases in smart buildings, factories, and critical infrastructures to ensure industrial applicability.
Based on available project data, specific licensing terms and IP ownership details are not mentioned.
The project focuses on the IoT-Edge-Cloud continuum, providing AIaaS and CaaS techniques for robust operation across these layers.
The project runs from 2024-04-01 to 2027-03-31, with the framework being tested in pilots during this period.
The project features a strong industrial lean with 14 industry partners (52% of the total 27 partners), including 6 SMEs. This high industry ratio, combined with 12 universities across 14 countries, suggests a high probability of commercial translation and practical application in real-world settings.
Contact ETHNICON METSOVION POLYTECHNION in Greece
Contact us to connect with the PANDORA consortium for pilot integration.
